A team of twelve may still have only one person who knows how to complete an essential monthly task. Headcount looks healthy; continuity is fragile. The “bus factor” is a bluntly named idea used in software work to discuss how concentrated critical knowledge has become.
JetBrains Research’s Bus Factor Explorer studies this problem through software-project information. Such models are useful within their assumptions, but a contribution history is not a complete map of what people know. Applying the idea to a wider team requires conversations about actual work, not a numerical ranking of employee value.
Find the task that stops
Ask: “If the usual person were unavailable next week, which task would we struggle to complete correctly?” This keeps the discussion about a process. It avoids treating expertise as a fault or demanding that a knowledgeable colleague document everything at once.
For an illustrative team, the fragile task might be preparing a supplier payment file. A folder of instructions is a start, but someone else also needs the appropriate access, context and opportunity to practise. Documentation without permission or practice can create false reassurance.
Test one handover while help is available
- Choose a recurring task with a clear outcome.
- Have the experienced person explain the decisions, exceptions and checks.
- Let another authorised colleague perform the task with supervision.
- Record where the handover depended on knowledge that had not been written down.
Use approved access arrangements; sharing a password is not a sound substitute for giving a colleague their own authorised access. Protect confidential material while making the legitimate workflow understandable.
Recognition matters too. The expert is contributing time and judgement to reduce a team risk. Include that work in priorities rather than adding it invisibly to an already full workload.
A useful outcome is modest and observable: another person can complete a specific process and knows when to ask for help. That is stronger evidence of resilience than declaring that every task has a backup name on a spreadsheet.